Perhaps the nearest approach thereto is a fifteenth-century Catalan map[208] (Fig. 7) preserved in the Ambrosian library in Milan, which as we have seen in Chapter IV, presents Greenland (Illa Verde) as a great elongated rectangle of land in northern waters, having a concave southern end. Below this, beyond a narrow interval of water, appears a large round island, the direction certainly calling for Labrador or Newfoundland, probably the latter. The minimizing of the distance between these land masses may indicate some report of the ease with which the crossing was effected. At any rate, unless we are prepared to set aside the testimony of the map altogether as mere fancy work, we must acknowledge that some one had a general impression of land in mass south or southwest of Greenland and reasonably accessible therefrom. BRAZIL ISLAND IN THE PLACE OF MARKLAND The name Brazil given to this island on the map and its disk-like form link it to the long series, already discussed, of “Brazil islands,” approximately in the latitude of Newfoundland, on the medieval maps, beginning with that of Dalorto of 1325[209] (Fig. 4). Usually, as in this last instance, they have the circular form--sometimes, however, being annular, with an island-studded lake or gulf inside, and sometimes being divided into two parts by a curved channel. Usually, too, the station of this Brazil is pretty near southern Ireland, off the Blaskets, but sometimes it is carried out into mid-Atlantic, and in the sixteenth-century maps of Nicolay[210] (1560; Fig. 6) and Zaltieri[211] (1566) it is taken clear across to the Banks of Newfoundland or a little nearer inshore. From various mutually corroborative indications, I have been impressed with the belief that it is probably a record of some early crossing of the Atlantic from Ireland; but whatever the explanation, Brazil Island remains one of the most interesting of map phenomena. Its name was somehow passed along to Terceira of the Azores, where there is still a Mt. Brazil, and long thereafter to the largest of South American countries. Its appearance near Greenland and as a substitute for Markland is not easily accounted for. The matter is indeed complicated on this fifteenth-century map by the appearance of a second Brazil (of the channeled type) in the middle of the Atlantic. It may be that the cartographer was familiar with this form and kind of presentation in older maps and did not feel warranted in giving up _that_ “Brazil;” but had received convincing information of lands southwest or south of Greenland, with some suggestion of Brazil as a name traditionally associated with such discoveries, and so drew and named it. Undoubtedly the map is the work of a man well acquainted with the first disk form of Brazil and the later channeled or divided form, beside having some knowledge of later discoveries in Greenland and beyond.